********************************************************************************
Title: Ambiguous Platforms and Correlated Preferences: Experimental Evidence
Authors: Juha Tolvanen, James Tremewan and Alexander K. Wagner
********************************************************************************


********************************************************************************
INTRODUCTION
********************************************************************************

This zip file contains the following items:
- Data
- Code (Do-Files)
- Ethics Certificate VCEE at U Vienna

Each of the above is described and detailed in turn.

********************************************************************************
 DATA
********************************************************************************

contains the following items:

final_dataset.dta


- Contains behavioral data of subjects in experiment used for analysis
- Those interested in recreating the exhibits in the paper use this dataset


********************************************************************************
CODE
********************************************************************************

contains the following items:

Do-File (Stata)

- For running the analyses included in the paper use the final dataset and set the path for the folder where this file is located. 
- The code files is commented and indicates which exhibits in the paper they generate. 
-  STATA16.1 was used for analysis.



********************************************************************************
ETHICAL STATEMENTS
********************************************************************************

contains the following items:

Ethics_Certificate_VCEE_experiment.pdf




********************************************************************************
 DATA DICTIONARY FOR FINAL DATASET
********************************************************************************

This is the data dictionary for the above final_code.do file. It describes the meaning of each of the key variables used in the analysis. Additional comments are found in the code file.


*** *** *** 
CANDIDATES
*** *** ***
 
Colours/Platforms:
1 = Grey; 2 = Black; 3 = White; 4 = Decide later
colorvoter:	1,2,3
colorpolitician:  1,2,3

platform_adj: 1 = Grey" 2 = Own colour 3 = Other colour 4 = Decide Later (also Non-commit, Ambiguous in the main text)
noncommit = 1 if platform is Decide Later
oplatform = Platform of opponent
mplatform = Matched Platform

Matched Platforms:
GBk=1;
GW=2;
GBb=3;
BkW=4;
BkBb=5;
WBb=6;
TIE=7;


*** *** 
VOTERS
*** ***

Baseline (BL treatment)
vote1: Grey vs Black (1 = vote for Black)
vote2: Grey vs White (1 = vote for White)
vote3: Grey vs Decide later (1 = vote for Decide later)
vote4: Black vs White (1 = vote for White)
vote5: Black vs Decide later (1 = vote for Decide later)
vote6: White vs Decide later (1 = vote for Decide later)

Incumbency (KC - known centrist treatment)
vote2: Grey vs Black (1 = vote for Black)
vote3: Grey vs White (1 = vote for White)
vote4: Grey vs Decide later (1 = vote for Decide later)

votenoncommit = 1 if voter votes for Decide Later over Grey

role=1 candidate, =2 voter


*** ****** ****** ***
ADDITIONAL VARIABLES (from additional tasks and questionnaire)
*** ****** ****** ***
Incumbent:
dummy for treatment: 	
=1 incumbent (Treatment: KC),
=0 baseline (Treatment: BL);

* strategic reasoner (SR) in the main text
Guess: Guess in beauty contest
level1plus: 1 if Guess < 34

* correlation awareness task:
Guess1: Guess in urn question with no correlation
Guess2: Guess in urn question with correlation

updateexact:  =0 correlation neglect, if =1 correlation aware (CA) in the main text

risk: unincentivised risk question (higher number implies seeing self as someone who takes more risk) from 0 to 10

